After Chatbot Grok Xai spent several hours on Wednesday, telling each user X who would listen to the fact that the claim of white genocide in South Africa is highly controversial, the company blamed this behavior for “unauthorized modification” of the GROK code.
In the Wednesday outbreak of the explosion, Grok put a discussion about the alleged white genocide in South Africa in various answers to X, regardless of the topic. GroK discussed the death of white farmers in response to a video from drinking water, he linked the song “Kill the Burss” with a question about Spongebob squarepantsAND Broke the problem in full patois. Even the general director of Opeli, Altman himself, took care of the action, making fun of the public division of Chatbot’s rival.
IN OX statement The company said that someone modified the AI bot system, “which directed the groc to a specific answer about the political.” This modification “has violated the internal principles and basic values of XAI”, and the company claims that “has carried out a thorough investigation” and implements recent funds to improve “transparency and reliability”.
These funds include publishing hints at the GROK level level publicly on githubLaunching the 24/7 monitoring team to perform such problems faster, and adding “additional controls and resources to make sure that XAI employees cannot modify the monitor without review.”
Xai already had this problem. In February, the company blamed the former nameless employee-openai for shifting the change in the hint of GROK, in which Chatbot disregarded all sources that accused Elon Musk or Donald Trump of the spread of disinformation. At that time Xai’s Head of Engineering, Igor Babuschkin, the employee said He was able to change “without asking anyone in the company for confirmation.”